I've given a first pass of Trac deploy via Ansible [1]
Now the apache config is also deployed via Ansible and we're again
having the problem of "single source of truth" because currently
the /etc/apache2 directory on the production host is a work dir
of a git repository which is a clone of a repo published via Gitea [2]
Right now I'm manually committing and pushing the changes deployed via
Ansible but it's an annoyance I'd like to avoid, especially since the
git repository is already out of date, missing a lot of files that
were removed in the past (like mod_evasive) and now present again,
failing to tell who added them when and why.
Another issue with the published git repository is that it fails
to deal with the configuration differences between production and
staging which are instead handled by the Ansible deployment via
templates and host variables.
Are you ok with removing the remote from the local git repo as
a start ? If not else it would reduce the odds of accidentally
pushing contending changes on the same branch between staging
and production.
[1] #2731 (deploy trac via ansible) – OSGeo
[2] https://git.osgeo.org/gitea/sac/tracsvn-apache-config.git
--strk;
Libre GIS consultant/developer
strk's services